常规Java课程思政资源个性化推荐系统的推荐效果不佳,因此提出基于深度学习的Java课程思政资源个性化推荐系统.首先设计资源存储器和资源处理器等系统硬件,其次基于深度学习算法构建个性化推荐模型,最后结合数据库及课程资源管理模块完成系统软件设计.测试结果表明,设计的系统能够实现课程资源的个性化推荐,推荐资源与用户需求资源之间的适配度更高.

针对常规教学资源共享系统存在响应时间较长和丢包率较大的问题,提出基于云计算的产学研模式教学智能共享系统.硬件设计方面,设计了共享信息微处理器和智能数字传感器;软件设计方面,在云计算技术的支持下,将接收的教学数据按照教学资源属性进行分级处理,确保教学资源的完整性,实现了资源共享.实验结果表明,该系统的响应时间更短,丢包率更低,具有更高的应用价值.

In order to improve the quality and effect of online education services in the new era and promote the virtuous circle development of online education service industry, an adaptive mobile learning service intelligent recommendation method based on deep learning algorithm is studied. Firstly, it analyzes the basic architecture of the intelligent recommendation method of adaptive mobile learning service, then discusses the collection and processing of mobile learning data information,and finally establishes the intelligent recommendation model based on deep learning, and tests and analyzes it. The experimental results show that the method is efficient and real-time, and can meet the basic requirements of actual projects for running speed and recommendation performance.

当前学习资源规模较大,导致分析难度较大,分析结果可靠性较低,为此,提出基于JavaEE的海量学习资源分析平台设计与实现研究.将具有强大存储功能的SDBPNPZ-256G-XI存储器和稳定转换功能的523-FCE17A15AD290适配器作为平台的硬件环境,利用JavaEE从存在需求角度对资源进行二级化处理,根据待分析资源的相似性和参数跨度实现对其的分析.测试结果:设计平台对资源所属类别的分析结果准确率可以达到98.8％,对于资源属性的分析结果准确率可以达到96.15％.

现有智能化运维应用在测试过程中抗噪声能力差、测试精度低、稳定性差,并且花费时间较长,影响测试效果,因此对网络数据中心IT设备人工智能化运维应用进行研究.通过VRNN算法定位网络数据中心IT设备中的异常数据,结合可量化的安全评估模型进行人工智能化的运维应用测试.通过实验能够证明,提出方法的抗噪声能力较好,测试精度最高能够达到92％以上,测试过程μzt值可达到0.39,稳定性较好,并且测试所需时间短,最快在60 s可以完成3000个数据的测试,说明提出方法具有较好的实用性.

计算机技术的发展推动了社会进步,程序设计类课程因此受到了重视.随着信息时代的到来,传统的程序设计教学方式已经无法满足目前的教学需求.基于此,设计程序设计类课程线上教学互动平台.硬件部分设计ARM处理器和S3C4510B储存芯片;软件部分,分析程序设计线上教学互动需求,设计程序设计线上教学功能模块,设计数据库,实现了线上教学互动.测试结果表明,设计的程序设计类教学平台性能良好,有一定的应用价值,可以作为后续程序设计类课程教学创新的参考.

海量网络内容中,许多话题结构相近、关系相似,传统的话题倾向型模型,只依赖一个固定特征词作为模型搜索指标,导致模型搜索时遗失部分相似性内容,因此构建具备相似性搜索能力的话题倾向型模型.该模型利用改进蚁群聚类算法,划分海量网络内容;采用时间序列法选择同类话题,追踪话题倾向;根据低秩矩阵和约束条件,计算倾向相似度,通过设置相似性搜索指标,实现模型的全局搜索.实验结果表明,与传统方法构建的模型相比,此次构建的模型以多个同类型的相似指标进行搜索,得到的数据更多,降低了相似性内容的遗失数量.

在通信系统中,连续时间滤波器是不可或缺的部分,如何灵活简洁地设计出易于集成、高频特性好、传输特性误差小的滤波器是电路与系统学界研究的重要方向.本文给出一种基于积分器模块的通用设计方法.该方法实现的电路结构中电阻与电容的元件数值合适,易于集成电路的工艺实现,并具有低的元件参数分散度.仿真结果表明所提出的电路方案正确有效,适于全集成.

借鉴已有颜色空间转换方法,通过认真分析RGB、YUV颜色空间模型,结合YUV采样格式和存储方式,本文提出RGB到YUV转换的功能需求,并给出了相应的转换公式.在此基础上,本文利用VC++技术将颜色空间转换思想加以实现,并详细描述了RGB到YUV转换的关键技术及核心算法.软件运行结果显示,该转换方式在视频格式处理方面具有明显的优势.

采用三维激光扫描技术再现识别激光全息三维图像时,未有效分类激光全息三维图像点云边界数据,存在图像再现识别过程耗时长的弊端,深入研究基于大数据分析技术的激光全息三维图像识别方法,采用大数据分析技术构建一种用于激光全息三维图像点云边界数据提取的云模型,采用该模型有效分类激光全息三维图像点云边界数据;采用激光全息扫描装置中得到分类的边界点云数据至激光全息图距离以及激光扫描光线的方位角等参数,依据这些参数获取激光全息三维图像中点云数据的三维坐标,对该坐标进行投射变换后于计算机上呈现立体视觉明显的激光全息三维图像,准确再现识别激光全息三维图像.实验结果说明,本文方法识别激光全息三维图像的用时少,识别激光全息三维图像的错误率低,是一种高效、可靠的激光全息三维图像识别方法.

We mainly introduce some related concepts in decision information system,such as discernible relation and relative discernible relation,and complete the attribute set independence judgment.Connecting the discernibility and relative discernibility of the attribute set with the object of identification of the property set,we also study the evaluation methods of attribute importance between the two kinds of relative identification and finish the judgment of attribute's independence or dependence and whether to do reduction of decision informa-tion system.Based on the relative recognition relation,the corresponding improvement algorithm is given,by which the variation of the relative recognition caused by the addition or subtraction of attribute set attributes is described.The algorithm determines the relatively dis-cernible relation of each attribute from the condition attribute set,taking the largest number of object as the reduction set where the attrib-utes are gradually added until meeting the conditions of reduction.It is a kind of seedless attribute reduction algorithm,which is reduced to a certain extent regardless of time complexity or reduction workload.Its validity is verified by an example.

Combustion supporting materials for indoor temperature change research, there has been the problem of inaccurate analysis, based on image intelligent judgment of indoor combustion supporting material temperature rising trend analysis method is proposed. By using digital image processing temperature intelligent detection methods, the introduction of resistance to high temperature charge coupled device (CCD) to obtain the real-time images of the interior materials, combustion temperature change, by colorimetric algorithm to extract interior materials, combustion temperature change image features, using BP neural network algorithm, indoor materials, combustion temperature change of image intelligent recognition. Experimental results show that the detection system can be used to extract the characteristic information of the extreme points of the temperature change, and to prevent the occurrence of fire.